INDEX

Explanations

phrases related to learning from experiences or mistakes

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize learning from experiences

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

yg

-0.77

hai

-0.74

merce

-0.74

stem

-0.73

rils

-0.73

important

-0.73

ifling

-0.73

idy

-0.72

aqu

-0.71

icide

-0.70

POSITIVE LOGITS

 afar

1.26

 whence

1.02

 scratch

0.90

 thence

0.81

 abroad

0.71

 Sasha

0.68

TR

0.65

 Dirk

0.62

 Drew

0.61

 Tony

0.61

Activations Density 0.084%

phrases related to learning from experiences or mistakes

phrases that emphasize learning from experiences

No Comments

No Known Activations