INDEX

Explanations

personal narratives containing details about decisions and experiences

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 WATCH

-0.76

ebin

-0.68

 HERE

-0.65

claimer

-0.64

osponsors

-0.62

 Appears

-0.62

Update

-0.61

 currently

-0.59

ogene

-0.59

zbollah

-0.59

POSITIVE LOGITS

 mattered

0.94

 lacked

0.88

 tended

0.87

 resembled

0.80

 seemed

0.79

 weren

0.77

 consisted

0.77

 knew

0.77

didn

0.76

 outnumbered

0.76

Activations Density 7.068%

personal narratives containing details about decisions and experiences

No Comments

No Known Activations