INDEX

Explanations

references to historians and scholarly discussions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

engers

-0.66

 Krypt

-0.65

 Delivery

-0.65

 blockers

-0.62

Interstitial

-0.62

 routed

-0.61

 Customers

-0.61

 speedy

-0.59

 takeoff

-0.59

 Braves

-0.59

POSITIVE LOGITS

 scholar

1.15

sonian

1.13

 scholars

1.07

 laureate

1.05

icist

1.04

itatively

1.04

 specializing

0.90

 historian

0.87

analy

0.85

Reviewed

0.83

Activations Density 0.090%

references to historians and scholarly discussions

No Comments

No Known Activations