INDEX

Explanations

words related to being accompanied or joined by someone or something

oai_token-act-pair · gpt-3.5-turbo

terms related to assistance and accompaniment

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

 Tribunal

-0.69

 datas

-0.65

 nuisance

-0.61

 mouth

-0.60

 Fiji

-0.60

 Combine

-0.59

 Paste

-0.59

knife

-0.59

 Netherlands

-0.59

 allowable

-0.58

POSITIVE LOGITS

by

0.97

By

0.91

by

0.89

ĸļ

0.83

Ń·

0.81

 aback

0.81

rored

0.79

BY

0.79

>]

0.78

By

0.76

Activations Density 0.282%

words related to being accompanied or joined by someone or something

terms related to assistance and accompaniment

No Comments

No Known Activations