INDEX

Explanations

references to news articles or reporters

oai_token-act-pair · gpt-3.5-turbo

possessive phrases indicating attribution to various sources or authors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

PLA

-0.82

Sov

-0.81

#$#$

-0.78

unin

-0.78

Ã©t

-0.77

$$$$

-0.76

oves

-0.74

utory

-0.71

RFC

-0.71

sov

-0.71

POSITIVE LOGITS

 Geoff

1.25

 Jonathan

1.22

 Jeffrey

1.22

 Jason

1.21

 Jennifer

1.19

 Andrew

1.19

 Matthew

1.19

Ian

1.19

 Erik

1.18

 Jesse

1.18

Activations Density 0.098%

references to news articles or reporters

possessive phrases indicating attribution to various sources or authors

No Comments

No Known Activations