INDEX

Explanations

occurrences of repeated letter patterns or alliteration

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

Õ

-0.71

Quantity

-0.64

 Scrolls

-0.61

Percent

-0.57

recomm

-0.56

¥µ

-0.55

plane

-0.54

\/\/

-0.53

 Witnesses

-0.53

 comprom

-0.53

POSITIVE LOGITS

rency

1.05

incial

0.78

acan

0.75

perse

0.72

atorial

0.72

cious

0.69

aca

0.68

ida

0.67

rane

0.67

zyme

0.66

Activations Density 0.021%

occurrences of repeated letter patterns or alliteration

No Comments

No Known Activations