INDEX

Explanations

phrases that indicate a growth or increase in phenomena

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

VICE

-0.72

chy

-0.66

MRI

-0.66

beit

-0.62

ãĥĩãĤ£

-0.62

cise

-0.61

ãĥĨãĤ£

-0.61

selves

-0.61

buff

-0.60

 throats

-0.60

POSITIVE LOGITS

ivist

0.82

ancy

0.77

 tide

0.77

rise

0.77

olithic

0.73

 tides

0.72

xual

0.70

down

0.69

ups

0.69

anqu

0.68

Activations Density 0.026%

phrases that indicate a growth or increase in phenomena

No Comments

No Known Activations