INDEX

Explanations

text related to significant discoveries or findings

oai_token-act-pair · gpt-3.5-turbo

statements related to scientific discoveries

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

annis

-0.88

overe

-0.83

Ã±a

-0.71

mph

-0.71

quer

-0.70

cemic

-0.70

attery

-0.68

 bailed

-0.67

oiler

-0.66

aded

-0.65

POSITIVE LOGITS

 Primordial

0.77

Ô

0.76

iveness

0.75

 discovery

0.73

ism

0.72

éļ

0.71

 discoveries

0.70

ivity

0.69

ä½¿

0.68

Flags

0.67

Activations Density 0.021%

text related to significant discoveries or findings

statements related to scientific discoveries

No Comments

No Known Activations