INDEX

Explanations

verbs or adverbs indicating actions or states occurring after a specific event

oai_token-act-pair · gpt-3.5-turbo

terms related to events and their consequences

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Â£

-1.92

 realise

-1.79

 recognise

-1.79

 organisation

-1.72

 Whilst

-1.70

 realised

-1.69

 recognised

-1.67

 colour

-1.66

 maths

-1.66

 organise

-1.66

POSITIVE LOGITS

colored

1.40

avored

1.30

defense

1.28

 publicized

1.23

 behaviors

1.22

 honoring

1.22

 canceled

1.20

imize

1.19

 favorably

1.18

 victimized

1.17

Activations Density 0.166%

verbs or adverbs indicating actions or states occurring after a specific event

terms related to events and their consequences

No Comments

No Known Activations