INDEX

Explanations

descriptions of ongoing activities or processes

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ns

-0.67

TY

-0.66

ptive

-0.64

ags

-0.63

apolis

-0.62

meg

-0.61

ulous

-0.60

zed

-0.60

oric

-0.60

aign

-0.58

POSITIVE LOGITS

side

1.08

wagon

0.78

 side

0.74

Side

0.73

 Aven

0.70

shore

0.69

isan

0.69

*=-

0.68

 Clause

0.66

 Came

0.66

Activations Density 1.139%

descriptions of ongoing activities or processes

No Comments

No Known Activations