INDEX

Explanations

phrases or sentences indicating a sequence of actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "then," indicating sequential or instructional contexts

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

ut

-0.66

pires

-0.65

orum

-0.64

chio

-0.63

cap

-0.63

ction

-0.61

borgh

-0.61

scription

-0.60

caps

-0.60

hered

-0.59

POSITIVE LOGITS

srf

0.82

 proceeded

0.81

Ń·

0.78

veland

0.77

Ļ½

0.77

EStream

0.75

 proceed

0.69

 Extras

0.68

————————————————

0.68

igslist

0.68

Activations Density 0.045%

phrases or sentences indicating a sequence of actions

instances of the word "then," indicating sequential or instructional contexts

No Comments

No Known Activations