INDEX

Explanations

phrases indicating emphasis or importance

oai_token-act-pair · gpt-3.5-turbo

phrases indicating continuity or progression in a discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

eers

-0.63

alis

-0.61

arettes

-0.57

 Newsp

-0.56

æĢ

-0.55

 Kens

-0.55

 Dragonbound

-0.55

arette

-0.54

urated

-0.54

free

-0.53

POSITIVE LOGITS

verning

0.93

 beyond

0.89

 overboard

0.87

lems

0.85

 nowhere

0.81

 unnoticed

0.80

 farther

0.77

 hand

0.76

 deeper

0.75

ggle

0.73

Activations Density 0.057%

phrases indicating emphasis or importance

phrases indicating continuity or progression in a discussion

No Comments

No Known Activations