INDEX

Explanations

non-English characters and symbols

oai_token-act-pair · gpt-3.5-turbo

sequences of non-standard characters or symbols

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 bonded

-0.86

 bour

-0.86

 biod

-0.81

 Brist

-0.76

 rigged

-0.74

 hoax

-0.72

 ranc

-0.72

 vitri

-0.72

 Mead

-0.71

 Bradford

-0.71

POSITIVE LOGITS

ãģĦ

2.19

ãģ

2.15

ãĤĭ

2.13

ãģŁ

2.13

ãĤ

2.12

ãģ¾

2.11

ãģĵ

2.07

ãĤĤ

2.07

ãģĭ

2.06

ãģĹ

2.06

Activations Density 0.016%

non-English characters and symbols

sequences of non-standard characters or symbols

No Comments

No Known Activations