INDEX

Explanations

numerical values, particularly those related to percentages and ratings

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

ophon

-0.79

ĸļ

-0.77

paio

-0.76

sein

-0.71

phis

-0.68

yle

-0.68

chio

-0.65

icago

-0.64

ichick

-0.64

ei

-0.63

POSITIVE LOGITS

th

0.97

isher

0.94

ishers

0.92

ishing

0.92

%-

0.91

0.89

0.88

%:

0.88

Activations Density 0.044%

numerical values, particularly those related to percentages and ratings

No Comments

No Known Activations