INDEX

Explanations

words related to providing support or evidence for a claim or action

oai_token-act-pair · gpt-3.5-turbo

phrases related to support or endorsement

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

anooga

-0.77

anny

-0.73

ities

-0.72

odor

-0.72

hester

-0.72

tein

-0.71

Osc

-0.67

itarian

-0.67

apeake

-0.67

awk

-0.67

POSITIVE LOGITS

swing

0.82

 backing

0.76

 backed

0.76

drive

0.74

abies

0.74

track

0.71

GROUND

0.70

backed

0.70

drops

0.68

country

0.67

Activations Density 0.016%

words related to providing support or evidence for a claim or action

phrases related to support or endorsement

No Comments

No Known Activations