INDEX

Explanations

words related to surprises or unexpected events

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "shock."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

•layer 1 - surprise

No Comments

Negative Logits

guyen

-0.86

amins

-0.84

uties

-0.76

 subp

-0.75

©¶æ

-0.74

orney

-0.72

ccording

-0.70

igion

-0.67

allery

-0.66

 Parenthood

-0.65

POSITIVE LOGITS

wave

1.06

waves

1.05

 shock

1.04

 absor

1.03

shock

0.92

 shocks

0.89

 Shock

0.86

Shock

0.85

imaru

0.81

ingly

0.81

Activations Density 0.009%

words related to surprises or unexpected events

mentions of the term "shock."

No Comments

No Known Activations