INDEX

Explanations

references to future plans or intentions

oai_token-act-pair · gpt-3.5-turbo

references to intentions or future arrangements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

********

-0.70

aque

-0.65

 Attribution

-0.65

ross

-0.63

IAS

-0.63

inguished

-0.62

Reading

-0.62

"]=>

-0.62

fry

-0.62

MAL

-0.61

POSITIVE LOGITS

 plans

3.88

 Plans

3.12

 plan

2.41

 Plan

1.80

Plan

1.79

 planned

1.76

 intentions

1.75

 proposals

1.73

plan

1.72

 ambitions

1.70

Activations Density 0.019%

references to future plans or intentions

references to intentions or future arrangements

No Comments

No Known Activations