INDEX

Explanations

references to different options or alternatives

oai_token-act-pair · gpt-3.5-turbo

conjunctions that connect contrasting or alternative ideas

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

ires

-0.72

efer

-0.71

onday

-0.67

ETS

-0.65

 Pony

-0.63

elson

-0.63

NOW

-0.63

Hoo

-0.63

legraph

-0.61

ocracy

-0.61

POSITIVE LOGITS

acle

1.38

nam

1.30

acles

1.28

chid

1.18

 otherwise

1.12

Else

1.11

chard

1.08

ifice

1.08

acular

1.05

nery

0.95

Activations Density 0.151%

references to different options or alternatives

conjunctions that connect contrasting or alternative ideas

No Comments

No Known Activations