INDEX

Explanations

references to the number "two" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to the number two 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 @ 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

IAS

-0.81

brance

-0.80

urden

-0.76

atown

-0.74

¶æ

-0.72

ISTORY

-0.70

imester

-0.70

displayText

-0.69

LEASE

-0.69

erver

-0.69

POSITIVE LOGITS

 distinct

1.20

 different

1.19

 separate

1.19

 disparate

1.10

 identical

1.05

 strangers

0.98

 contrasting

0.97

 dudes

0.96

 halves

0.96

 pairs

0.95

Activations Density 0.192%

references to the number "two" in various contexts

references to the number two in various contexts

No Comments

No Known Activations