INDEX

Explanations

abbreviated terms and references to locations

oai_token-act-pair · gpt-3.5-turbo

punctuation, specifically commas

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 thous

-0.73

coni

-0.65

 corrid

-0.64

 challeng

-0.63

ãĤ´ãĥ³

-0.63

 disadvant

-0.60

ithub

-0.60

 misunder

-0.59

 predec

-0.59

 behavi

-0.58

POSITIVE LOGITS

huh

0.71

 please

0.68

 however

0.68

 albeit

0.68

 including

0.61

 though

0.59

 meanwhile

0.59

 which

0.58

eh

0.55

LLC

0.54

Activations Density 0.448%

abbreviated terms and references to locations

punctuation, specifically commas

No Comments

No Known Activations