INDEX

Explanations

references to actions involving personal interactions or conflicts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

selves

-0.83

 husbands

-0.66

 collectively

-0.66

atures

-0.65

 unison

-0.64

hub

-0.61

 husband

-0.61

 womb

-0.61

 ballots

-0.60

 merger

-0.60

POSITIVE LOGITS

 himself

1.48

his

0.94

 Himself

0.93

panic

0.75

 erection

0.73

 girlfriend

0.69

Jr

0.66

Jr

0.66

 ejac

0.65

sul

0.64

Activations Density 0.709%

references to actions involving personal interactions or conflicts

No Comments

No Known Activations