INDEX

Explanations

mentions of specific individuals, primarily actors or characters in entertainment contexts

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

choice

-0.80

ERAL

-0.77

areth

-0.74

ulate

-0.74

ulating

-0.73

ablishment

-0.72

emonic

-0.72

owship

-0.72

ulates

-0.71

erman

-0.70

POSITIVE LOGITS

nton

0.87

 Bros

0.67

 Russo

0.66

icum

0.66

eru

0.66

 Thornton

0.64

 Weasley

0.64

 sidx

0.64

Smy

0.63

bably

0.63

Activations Density 0.007%

mentions of specific individuals, primarily actors or characters in entertainment contexts

No Comments

No Known Activations