INDEX

Explanations

phrases related to instances or individuals with high public visibility or importance

oai_token-act-pair · gpt-3.5-turbo

references to high-profile individuals or events

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

heit

-0.87

©¶æ¥µ

-0.83

shall

-0.81

ensibly

-0.75

raph

-0.73

ope

-0.71

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.71

opes

-0.70

abi

-0.69

anguage

-0.68

POSITIVE LOGITS

 endorsements

1.02

 endors

1.02

 celebrity

0.99

 endorsement

0.95

 bearer

0.91

 departures

0.82

 venture

0.81

 casualty

0.80

 blockbuster

0.79

 sporting

0.79

Activations Density 0.066%

phrases related to instances or individuals with high public visibility or importance

references to high-profile individuals or events

No Comments

No Known Activations