INDEX

Explanations

proper nouns, particularly names of individuals and titles

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

 supers

-0.67

CLS

-0.67

theless

-0.66

 Sakuya

-0.65

 Dire

-0.64

 Coastal

-0.62

GST

-0.62

 horizont

-0.61

ACC

-0.61

 USPS

-0.61

POSITIVE LOGITS

acci

0.95

aney

0.93

rower

0.93

beck

0.93

kowski

0.90

isner

0.90

riott

0.89

rage

0.89

ansky

0.88

zinski

0.88

Activations Density 0.077%

proper nouns, particularly names of individuals and titles

No Comments

No Known Activations