INDEX

Explanations

words related to control and power

oai_token-act-pair · gpt-3.5-turbo

instances of the word "control" 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

igmat

-0.70

udget

-0.62

rouse

-0.62

eday

-0.57

 Zeit

-0.56

rehend

-0.56

enegger

-0.56

GGGGGGGG

-0.55

 Lect

-0.55

sung

-0.55

POSITIVE LOGITS

 over

1.16

over

0.98

of

0.92

orship

0.88

overs

0.87

 OVER

0.82

 thereof

0.82

ignty

0.81

holding

0.80

 exercised

0.78

Activations Density 0.089%

words related to control and power

instances of the word "control" in various contexts

No Comments

No Known Activations