INDEX

Explanations

words that suggest intensity, importance, or high emotion

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dramatic" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ĥ¬

-0.83

mun

-0.80

spr

-0.78

lected

-0.75

claimer

-0.74

respect

-0.71

checked

-0.70

ledge

-0.70

pel

-0.70

jack

-0.69

POSITIVE LOGITS

 dramatic

0.93

 earthqu

0.90

 leaps

0.87

 reversal

0.83

 outburst

0.81

 highs

0.78

 unbeliev

0.77

 contrasts

0.77

 brill

0.76

 rendition

0.75

Activations Density 0.013%

words that suggest intensity, importance, or high emotion

instances of the word "dramatic" in various contexts

No Comments

No Known Activations