INDEX

Explanations

phrases indicating special emphasis or importance

oai_token-act-pair · gpt-3.5-turbo

emphasis on specific topics or attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ylon

-0.68

ences

-0.66

 essentially

-0.65

CT

-0.64

ensibly

-0.63

ruary

-0.63

 substitutes

-0.63

only

-0.62

 Anarchy

-0.62

offic

-0.61

POSITIVE LOGITS

 egregious

1.09

 noteworthy

1.04

 suited

0.93

 notable

0.90

 troublesome

0.90

 noticeable

0.88

 susceptible

0.85

 poignant

0.84

 advantageous

0.82

 vulnerable

0.82

Activations Density 0.050%

phrases indicating special emphasis or importance

emphasis on specific topics or attributes

No Comments

No Known Activations