INDEX

Explanations

mentions of specific terms related to the company or brand "Clear Channel" or specific locations such as "Concordia"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Clear" and related terms indicating clarity or clearance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ITAL

-0.87

FAULT

-0.78

LAB

-0.74

Bey

-0.73

tackle

-0.72

SN

-0.71

externalActionCode

-0.69

minecraft

-0.69

seless

-0.67

hed

-0.67

POSITIVE LOGITS

ance

1.01

 Clear

0.97

 Seal

0.87

Clear

0.81

ances

0.79

 Flags

0.78

opsis

0.77

Flag

0.76

 Triangle

0.76

 Conce

0.75

Activations Density 0.021%

mentions of specific terms related to the company or brand "Clear Channel" or specific locations such as "Concordia"

occurrences of the word "Clear" and related terms indicating clarity or clearance

No Comments

No Known Activations