INDEX

Explanations

words related to observation or analysis

oai_token-act-pair · gpt-3.5-turbo

phrases indicating perception or understanding

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

rontal

-0.76

rig

-0.68

 Died

-0.66

 Niet

-0.64

cu

-0.63

calling

-0.62

istries

-0.62

rang

-0.61

jab

-0.60

seless

-0.60

POSITIVE LOGITS

IDENT

0.76

ANC

0.72

arten

0.67

Tip

0.65

ANCE

0.63

udos

0.63

į

0.61

Footnote

0.60

Written

0.60

 evidenced

0.59

Activations Density 0.064%

words related to observation or analysis

phrases indicating perception or understanding

No Comments

No Known Activations