INDEX

Explanations

references to a character named "Doc"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Doc" or similar variations in the document

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

places

-0.71

 Admir

-0.61

Osw

-0.61

eering

-0.60

boarding

-0.60

 Virgin

-0.59

 parting

-0.59

 Rivals

-0.59

committee

-0.58

cffffcc

-0.57

POSITIVE LOGITS

herty

1.25

uments

1.15

ents

1.00

ilt

0.94

iencies

0.93

umen

0.88

ual

0.86

ilic

0.85

endum

0.84

sonian

0.83

Activations Density 0.037%

references to a character named "Doc"

occurrences of the word "Doc" or similar variations in the document

No Comments

No Known Activations