INDEX

Explanations

phrases indicating unexpected or surprising outcomes

oai_token-act-pair · gpt-3.5-turbo

phrases indicating realizations or conclusions about situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

icipated

-0.86

anwhile

-0.77

shaw

-0.72

odox

-0.72

attled

-0.70

tymology

-0.69

cius

-0.69

raltar

-0.69

arta

-0.68

cies

-0.68

POSITIVE LOGITS

to

0.71

 raining

0.68

Ñĭ

0.63

 that

0.62

wards

0.61

âĶĢ

0.58

 liking

0.58

 disadvant

0.57

 differently

0.57

 innoc

0.56

Activations Density 0.059%

phrases indicating unexpected or surprising outcomes

phrases indicating realizations or conclusions about situations

No Comments

No Known Activations