INDEX

Explanations

concepts related to truth, legitimacy, and moral lessons

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

FY

-0.67

 Verb

-0.62

oxide

-0.59

chy

-0.59

 Chains

-0.59

 stormed

-0.58

 totaled

-0.57

 Drunk

-0.57

catentry

-0.57

rive

-0.57

POSITIVE LOGITS

 inherent

0.79

 lurking

0.78

 parallels

0.77

 similarities

0.76

 precedent

0.75

paralle

0.73

 overlap

0.72

 underlying

0.71

 downside

0.70

hurst

0.69

Activations Density 0.093%

concepts related to truth, legitimacy, and moral lessons

No Comments

No Known Activations