INDEX

Explanations

terms associated with credibility and incredulity in statements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

meal

-0.74

hap

-0.71

 Abandon

-0.65

 pity

-0.64

ĵĺ

-0.63

 Hots

-0.63

Won

-0.62

assetsadobe

-0.61

 Feature

-0.60

 combatants

-0.58

POSITIVE LOGITS

ibly

1.47

enza

1.41

ulous

1.41

ulously

1.40

ence

1.23

ences

1.21

itious

1.16

iting

1.16

encing

1.15

ulent

1.13

Activations Density 0.002%

terms associated with credibility and incredulity in statements

No Comments

No Known Activations