INDEX

Explanations

letter 'u' followed by numbers, possibly related to referencing different versions or editions

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

 Attribution

-0.73

 Cosponsors

-0.72

 Payton

-0.64

*/(

-0.63

 horizont

-0.61

 skelet

-0.58

Ö¼

-0.58

 Drift

-0.58

 SERV

-0.56

 Emin

-0.56

POSITIVE LOGITS

cca

1.30

pport

1.30

ccess

1.22

isine

1.20

seless

1.19

cci

1.11

pta

1.07

itton

1.05

selage

1.01

pping

0.98

Activations Density 0.668%

letter 'u' followed by numbers, possibly related to referencing different versions or editions

No Comments

No Known Activations