INDEX

Explanations

mentions of someone not engaging or participating in a certain activity or scenario

oai_token-act-pair · gpt-3.5-turbo

instances of the word "been" in various contexts

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

still

-0.68

Days

-0.64

onomy

-0.63

 Graves

-0.61

 limbo

-0.60

now

-0.59

 aspiration

-0.58

 juices

-0.56

 Potential

-0.56

 Begin

-0.56

POSITIVE LOGITS

 able

0.97

 touched

0.88

 bothered

0.87

 tested

0.87

 harmed

0.86

 adequately

0.86

epad

0.82

 prosecuted

0.81

 properly

0.81

 satisf

0.81

Activations Density 0.073%

mentions of someone not engaging or participating in a certain activity or scenario

instances of the word "been" in various contexts

No Comments

No Known Activations