INDEX

Explanations

phrases indicating uncertainty or open-ended possibilities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating uncertainty or speculation about outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

sted

-0.78

zar

-0.77

arie

-0.76

edia

-0.76

ahime

-0.72

ullivan

-0.71

trak

-0.68

obile

-0.67

Ã©s

-0.66

odore

-0.64

POSITIVE LOGITS

 whatsoever

0.93

 imaginable

0.86

 occasion

0.79

pires

0.76

 loaf

0.72

ifice

0.71

 circumstance

0.71

 deems

0.70

 deem

0.69

 chooses

0.68

Activations Density 0.127%

phrases indicating uncertainty or open-ended possibilities

phrases indicating uncertainty or speculation about outcomes

No Comments

No Known Activations