INDEX

Explanations

phrases indicating opinions or statistics from survey respondents

oai_token-act-pair · gpt-3.5-turbo

references to survey participants or respondents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Sabb

-0.70

 Sabha

-0.69

cop

-0.67

 Tours

-0.64

 Tour

-0.63

 Herald

-0.63

 Cooking

-0.61

 Sale

-0.61

 slic

-0.61

rip

-0.61

POSITIVE LOGITS

 respondents

0.95

 respondent

0.88

uggest

0.79

 answ

0.75

 interviewed

0.75

hip

0.74

 surveyed

0.72

 undecided

0.70

 answered

0.69

 answering

0.68

Activations Density 0.019%

phrases indicating opinions or statistics from survey respondents

references to survey participants or respondents

No Comments

No Known Activations