INDEX

Explanations

Keywords related to coding such as 'get', 'set', and 'subscript'

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases centered around the verb "get" 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 @ 12-res-jb

Configuration

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

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_post

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

 Examiner

-0.61

 hears

-0.60

 barring

-0.59

 contemplation

-0.58

 Continent

-0.56

 prophet

-0.55

 deduct

-0.55

 mitigating

-0.54

 Â¶

-0.53

 Banks

-0.53

POSITIVE LOGITS

cloneembedreportprint

1.20

TING

1.16

tin

1.13

chell

1.06

rid

1.05

away

1.04

bent

1.00

ter

1.00

ters

0.94

started

0.94

Activations Density 0.045%

Keywords related to coding such as 'get', 'set', and 'subscript'

phrases centered around the verb "get" in various contexts

No Comments

No Known Activations