INDEX

Explanations

references to rankings or positions within a competitive 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ajor

-0.73

 Relations

-0.60

 Daylight

-0.60

Äĩ

-0.59

rain

-0.58

iasco

-0.56

rw

-0.56

itt

-0.56

Ago

-0.55

ripp

-0.55

POSITIVE LOGITS

 tier

1.06

most

0.99

ographical

0.92

 scorer

0.92

 contenders

0.92

ography

0.89

liest

0.89

eka

0.86

deck

0.85

 notch

0.84

Activations Density 0.028%

references to rankings or positions within a competitive context

No Comments

No Known Activations