INDEX

Explanations

references to research findings or investigation outcomes

oai_token-act-pair · gpt-3.5-turbo

phrases mentioning the results or outcomes of investigations, trials, or elections

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

Ã±o

-0.72

ener

-0.72

horn

-0.68

anty

-0.68

Rew

-0.66

GREEN

-0.66

ÃŁ

-0.66

oster

-0.64

ept

-0.64

arten

-0.62

POSITIVE LOGITS

 deed

0.65

 preliminary

0.63

Roe

0.61

©¶æ

0.60

ãĥĢ

0.59

ursday

0.59

Ĥ¬

0.58

 RELE

0.57

 disse

0.57

{}

0.56

Activations Density 0.157%

references to research findings or investigation outcomes

phrases mentioning the results or outcomes of investigations, trials, or elections

No Comments

No Known Activations