INDEX

Explanations

words related to historical figures, especially with titles like "Pope" or "Emperor."

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_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

izable

-0.96

izational

-0.87

ished

-0.86

ding

-0.85

ized

-0.83

ishment

-0.78

izers

-0.76

agements

-0.76

ishing

-0.76

izations

-0.76

POSITIVE LOGITS

 Caesar

1.07

XM

0.92

XII

0.88

 Malfoy

0.84

Pil

0.84

 Severus

0.80

cles

0.80

aurus

0.79

hip

0.78

 Claud

0.78

Activations Density 0.043%

words related to historical figures, especially with titles like "Pope" or "Emperor."

No Comments

No Known Activations