INDEX

Explanations

phrases related to legal actions or policy discussions

oai_token-act-pair · gpt-3.5-turbo

negative political or legal implications related to health care and welfare issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Dickinson

-0.79

ulhu

-0.72

 Frost

-0.71

 Indigo

-0.70

 Tempest

-0.69

 EntityItem

-0.69

 Guru

-0.68

 Paradox

-0.68

 Polo

-0.67

 Tolkien

-0.67

POSITIVE LOGITS

based

1.30

government

1.14

sized

1.13

related

1.12

oriented

1.08

agency

1.08

level

1.07

themed

1.07

focused

1.06

friendly

1.05

Activations Density 0.478%

phrases related to legal actions or policy discussions

negative political or legal implications related to health care and welfare issues

No Comments

No Known Activations