INDEX

Explanations

phrases or statements with strong assertiveness or emphasis

oai_token-act-pair · gpt-3.5-turbo

instances of the word "it" and variations indicating a statement or assertion

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

Mesh

-0.60

swing

-0.60

illas

-0.59

rame

-0.59

iggurat

-0.59

reet

-0.59

awareness

-0.58

atars

-0.57

imum

-0.56

ilated

-0.55

POSITIVE LOGITS

nor

1.22

nor

0.97

yet

0.89

 anymore

0.83

unless

0.81

Nor

0.76

atures

0.71

 merely

0.70

 nevertheless

0.66

Ļ

0.65

Activations Density 0.636%

phrases or statements with strong assertiveness or emphasis

instances of the word "it" and variations indicating a statement or assertion

No Comments

No Known Activations