INDEX

Explanations

commands or instructions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "command."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Democr

-0.78

 Flavoring

-0.70

"],"

-0.66

 Anita

-0.66

 Rounds

-0.65

osponsors

-0.63

Versions

-0.62

 Wheat

-0.62

Disclaimer

-0.61

POSITIVE LOGITS

 command

1.13

eering

1.10

 commands

1.02

command

1.01

 commanded

0.95

Command

0.87

 Command

0.84

eers

0.83

boats

0.80

 commanding

0.79

Activations Density 0.007%

commands or instructions

occurrences of the word "command."

No Comments

No Known Activations