INDEX

Explanations

occurrences of the word "Guggenheim."

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

andro

-0.67

 scenes

-0.64

 grids

-0.63

 circles

-0.63

 differential

-0.61

Cir

-0.61

 licens

-0.61

BAD

-0.60

 perspective

-0.60

 balcony

-0.59

POSITIVE LOGITS

vich

0.85

uru

0.82

uay

0.80

jee

0.79

iru

0.77

orthy

0.77

 Haku

0.72

anski

0.71

ucket

0.71

iple

0.71

Activations Density 0.061%

occurrences of the word "Guggenheim."

No Comments

No Known Activations