INDEX

Explanations

proper nouns referring to people or organizations

oai_token-act-pair · gpt-3.5-turbo

references to individuals or entities in health-related 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

shift

-0.66

 unbeliev

-0.57

 Principle

-0.56

venge

-0.56

sense

-0.56

grow

-0.55

 detection

-0.55

iscovery

-0.54

 realization

-0.54

union

-0.52

POSITIVE LOGITS

did

1.47

 didn

1.24

 gave

1.15

 took

1.14

 waited

1.14

DID

1.13

 didnt

1.12

 stayed

1.12

 went

1.10

 withdrew

1.09

Activations Density 0.366%

proper nouns referring to people or organizations

references to individuals or entities in health-related contexts

No Comments

No Known Activations