INDEX

Explanations

phrases indicating strong opinions or evaluations

oai_token-act-pair · gpt-3.5-turbo

statements about problems or conditions that lead to significant consequences

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

FN

-0.65

known

-0.61

SD

-0.60

zig

-0.60

pring

-0.59

than

-0.59

odd

-0.59

alias

-0.58

episode

-0.58

wig

-0.57

POSITIVE LOGITS

 deserves

1.45

 attracts

1.24

 evolves

1.19

 requires

1.19

 belongs

1.18

 seeks

1.16

 needs

1.15

 inspires

1.14

 strives

1.14

 relies

1.14

Activations Density 0.183%

phrases indicating strong opinions or evaluations

statements about problems or conditions that lead to significant consequences

No Comments

No Known Activations