INDEX

Explanations

the word "surprise" at various levels of activation

oai_token-act-pair · gpt-3.5-turbo

expressions of surprise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

nan

-0.83

folios

-0.81

©¶æ

-0.81

oreal

-0.81

odynam

-0.80

agra

-0.80

tein

-0.78

İĭ

-0.76

asus

-0.74

uel

-0.73

POSITIVE LOGITS

 surprise

0.81

 absor

0.79

 surprises

0.79

 guests

0.79

 Surprise

0.76

ingly

0.76

 Flavoring

0.75

 visitor

0.74

 Squid

0.74

ãģį

0.71

Activations Density 0.047%

the word "surprise" at various levels of activation

expressions of surprise

No Comments

No Known Activations