INDEX

Explanations

repeated mentions of specific names, likely related to a central character or figure in the context

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

arist

-0.83

addons

-0.78

resses

-0.71

adr

-0.69

arians

-0.69

ournals

-0.67

illance

-0.67

ributes

-0.67

romy

-0.66

saf

-0.66

POSITIVE LOGITS

 Holt

1.05

 Kinn

0.81

 Lester

0.80

 Browne

0.78

 Rouge

0.76

Pod

0.73

 Madd

0.72

 Hodg

0.72

 Kers

0.70

 Cain

0.70

Activations Density 0.004%

repeated mentions of specific names, likely related to a central character or figure in the context

No Comments

No Known Activations