INDEX

Explanations

words related to specific cultural or religious references

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

ding

-0.78

sburgh

-0.78

olson

-0.76

etsy

-0.70

assed

-0.67

intosh

-0.66

raltar

-0.65

lished

-0.64

iverpool

-0.64

igree

-0.64

POSITIVE LOGITS

Dum

0.73

qi

0.73

Tao

0.72

 Sabha

0.72

verse

0.72

plin

0.71

ze

0.71

iste

0.70

efully

0.70

jin

0.69

Activations Density 0.005%

words related to specific cultural or religious references

No Comments

No Known Activations