INDEX

Explanations

the word "premium" with varied activations

oai_token-act-pair · gpt-3.5-turbo

references to "premium" as a descriptor for various products or services

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

CLE

-0.76

prototype

-0.75

wright

-0.74

ADRA

-0.68

things

-0.66

İĭ

-0.65

mable

-0.64

phant

-0.63

arming

-0.63

iverpool

-0.62

POSITIVE LOGITS

cedes

0.85

iator

0.79

raviolet

0.76

 premium

0.73

 upkeep

0.72

 billing

0.69

endiary

0.68

 income

0.68

ately

0.68

ensable

0.67

Activations Density 0.029%

the word "premium" with varied activations

references to "premium" as a descriptor for various products or services

No Comments

No Known Activations