INDEX

Explanations

phrases indicating additional information or supplementary details

oai_token-act-pair · gpt-3.5-turbo

the word "Additionally" and its variations, indicating the presence of supplementary information

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

-0.80

 hearts

-0.68

venge

-0.68

enders

-0.65

 Fare

-0.64

fell

-0.63

ger

-0.62

liest

-0.61

ender

-0.60

orno

-0.59

POSITIVE LOGITS

 guiActiveUn

0.78

 importantly

0.72

æ©Ł

0.72

 noteworthy

0.69

ilitary

0.66

ally

0.65

notation

0.65

ificantly

0.64

yss

0.64

orthy

0.62

Activations Density 0.015%

phrases indicating additional information or supplementary details

the word "Additionally" and its variations, indicating the presence of supplementary information

No Comments

No Known Activations