INDEX

Explanations

parts of the text with emphasis markers

oai_token-act-pair · gpt-3.5-turbo

instances of the word "emphasis" and related phrasing

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

tre

-0.72

izons

-0.71

 Rebell

-0.68

STON

-0.68

wagen

-0.68

tein

-0.67

met

-0.66

 Skydragon

-0.66

Nazis

-0.65

apons

-0.63

POSITIVE LOGITS

 emphasis

1.05

phasis

0.93

xual

0.78

 shifts

0.78

inction

0.76

 reliance

0.76

estinal

0.75

uality

0.75

baugh

0.73

erences

0.73

Activations Density 0.018%

parts of the text with emphasis markers

instances of the word "emphasis" and related phrasing

No Comments

No Known Activations