INDEX

Explanations

phrases related to a specific group of people or identity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ans," indicating a focus on references to answers or responses

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

kefeller

-0.89

 WATCHED

-0.76

ptive

-0.76

ADS

-0.74

bable

-0.73

UTERS

-0.72

PRESS

-0.70

ô

-0.69

 buds

-0.67

 Jarrett

-0.67

POSITIVE LOGITS

igans

1.11

alis

1.04

hee

0.99

hip

0.96

haw

0.95

igan

0.93

laughter

0.90

omal

0.90

uman

0.89

hin

0.88

Activations Density 0.012%

phrases related to a specific group of people or identity

instances of the word "ans," indicating a focus on references to answers or responses

No Comments

No Known Activations