INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

issance

-0.68

 groom

-0.63

bered

-0.62

FORMATION

-0.60

QUEST

-0.60

cyl

-0.59

fixes

-0.58

 suppress

-0.58

ederation

-0.57

 Sparkle

-0.57

POSITIVE LOGITS

onom

1.03

ine

0.98

alo

0.92

ians

0.90

ulas

0.88

iders

0.86

ide

0.86

iera

0.85

ian

0.84

idal

0.82

Activations Density 3.004%

names of individuals

No Comments

No Known Activations