INDEX

Explanations

the word "repeat" and phrases indicating something happening again

oai_token-act-pair · gpt-3.5-turbo

instances of the word "repeat"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

sie

-0.76

uana

-0.74

ffe

-0.69

icial

-0.68

erie

-0.68

esthetic

-0.67

tesy

-0.66

avery

-0.66

ï¸

-0.66

otropic

-0.65

POSITIVE LOGITS

 repeat

0.92

 repeats

0.92

 playthrough

0.89

repeat

0.88

 offenders

0.83

 repeating

0.83

 repetition

0.81

 offender

0.80

lly

0.79

 occurrences

0.73

Activations Density 0.031%

the word "repeat" and phrases indicating something happening again

instances of the word "repeat"

No Comments

No Known Activations