INDEX

Explanations

instances of the word "let" and its variations, indicating instructions or commands

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

natureconservancy

-0.80

Interstitial

-0.76

 Donation

-0.67

blind

-0.66

emetery

-0.63

ombat

-0.63

sac

-0.62

cler

-0.62

hunt

-0.61

eworld

-0.60

POSITIVE LOGITS

itia

0.89

tering

0.81

imes

0.79

emort

0.77

icia

0.73

arations

0.73

 assume

0.72

tered

0.70

ters

0.68

uge

0.68

Activations Density 0.013%

instances of the word "let" and its variations, indicating instructions or commands

No Comments

No Known Activations