INDEX

Explanations

instances where an assumption is made or expected

oai_token-act-pair · gpt-3.5-turbo

instances of the word "assume"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

psey

-0.72

cgi

-0.71

HCR

-0.68

Sche

-0.65

encer

-0.63

 datas

-0.63

nels

-0.63

gom

-0.63

emetery

-0.62

History

-0.61

POSITIVE LOGITS

 assume

0.92

 assumed

0.79

 assuming

0.78

 assumes

0.70

IRT

0.66

 assumption

0.65

 presume

0.65

LY

0.64

urances

0.64

UME

0.64

Activations Density 0.014%

instances where an assumption is made or expected

instances of the word "assume"

No Comments

No Known Activations