INDEX

Explanations

phrases related to entry and exit actions

oai_token-act-pair · gpt-3.5-turbo

conjunctions and phrases indicating combinations or connections

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

swick

-0.73

conom

-0.71

ãĥ¯

-0.68

omorphic

-0.68

neys

-0.66

IX

-0.66

xxxxxxxx

-0.64

major

-0.62

xff

-0.61

Bron

-0.61

POSITIVE LOGITS

 departures

0.97

 outgoing

0.94

 subtract

0.91

 exit

0.88

 departure

0.85

 exits

0.80

exit

0.76

Remove

0.75

 depart

0.75

aft

0.73

Activations Density 0.159%

phrases related to entry and exit actions

conjunctions and phrases indicating combinations or connections

No Comments

No Known Activations