INDEX

Explanations

references to specific names and terms related to magic, mythology, and notable figures

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

teen

-0.71

 wasteland

-0.69

doms

-0.66

 lone

-0.64

venants

-0.63

eering

-0.62

eking

-0.61

ependent

-0.61

ez

-0.61

erness

-0.61

POSITIVE LOGITS

ited

0.95

ocrates

0.81

ENSE

0.80

iencies

0.80

alities

0.78

atically

0.77

ilk

0.74

icro

0.73

stre

0.72

iliar

0.72

Activations Density 0.009%

references to specific names and terms related to magic, mythology, and notable figures

No Comments

No Known Activations