INDEX

Explanations

phrases related to legal definitions, regulations, and criteria

oai_token-act-pair · gpt-3.5-turbo

phrases related to legal definitions and gender-related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 brass

-0.71

 snapped

-0.69

jer

-0.68

 stash

-0.66

 buddies

-0.65

 Geral

-0.65

 Lyft

-0.65

 bartender

-0.65

 iPhones

-0.65

 Winc

-0.64

POSITIVE LOGITS

 impair

1.06

adequ

0.96

reasonable

0.95

 impairment

0.93

sufficient

0.90

worthiness

0.90

ocative

0.88

reason

0.88

cellence

0.87

acceptable

0.87

Activations Density 0.721%

phrases related to legal definitions, regulations, and criteria

phrases related to legal definitions and gender-related concepts

No Comments

No Known Activations