INDEX

Explanations

references to telephone-related words

oai_token-act-pair · gpt-3.5-turbo

references to the telephone and telephone-related concepts

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

ebook

-0.86

 Ridley

-0.76

reci

-0.75

ean

-0.73

iverpool

-0.72

ded

-0.72

arent

-0.71

loe

-0.71

ellen

-0.71

iesta

-0.69

POSITIVE LOGITS

 booths

0.89

 telephone

0.87

 handset

0.85

 answ

0.83

 Telephone

0.83

 poles

0.83

 booth

0.82

 communication

0.79

NetMessage

0.77

 operators

0.77

Activations Density 0.016%

references to telephone-related words

references to the telephone and telephone-related concepts

No Comments

No Known Activations