INDEX

Explanations

mentions of specific individuals' names and terms associated with microwaves and decomposition

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

WAYS

-0.74

 Clockwork

-0.72

tips

-0.70

tenance

-0.68

 Colonial

-0.66

Interstitial

-0.65

 cycle

-0.65

 Gemini

-0.64

Magikarp

-0.64

tarian

-0.63

POSITIVE LOGITS

yre

1.30

osh

1.17

eneg

0.98

ott

0.96

ire

0.95

iet

0.94

orm

0.92

athy

0.92

olly

0.92

ermott

0.92

Activations Density 0.005%

mentions of specific individuals' names and terms associated with microwaves and decomposition

No Comments

No Known Activations