INDEX

Explanations

references to news sources or publishers

oai_token-act-pair · gpt-3.5-turbo

possessive forms related to various entities or organizations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

#$#$

-0.80

$$$$

-0.79

Ã©t

-0.76

PLA

-0.74

Ùĩ

-0.74

those

-0.73

\-

-0.72

Ø§

-0.72

ET

-0.72

},

-0.70

POSITIVE LOGITS

 newest

1.04

 Kevin

1.00

 Brian

0.97

 Erik

0.95

 Darren

0.95

 Jeffrey

0.94

Ian

0.94

 Josh

0.94

 Geoff

0.93

 chief

0.93

Activations Density 0.131%

references to news sources or publishers

possessive forms related to various entities or organizations

No Comments

No Known Activations