INDEX

Explanations

words or phrases that exhibit strong emotional reactions or significant implications, especially in political or social 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icult

-0.74

ording

-0.67

aimon

-0.66

oted

-0.65

iewicz

-0.65

oids

-0.65

assian

-0.64

utterstock

-0.64

icultural

-0.63

emonic

-0.62

POSITIVE LOGITS

vre

1.15

¬

0.99

lette

0.89

·

0.88

tis

0.85

sin

0.84

ÃįÃį

0.82

0.79

¹

0.79

¸

0.76

Activations Density 0.003%

words or phrases that exhibit strong emotional reactions or significant implications, especially in political or social contexts

No Comments

No Known Activations