INDEX

Explanations

places or buildings with historical significance

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

mie

-0.78

schild

-0.75

nir

-0.74

ELL

-0.71

CAST

-0.68

lessly

-0.67

feed

-0.67

REE

-0.66

rieg

-0.65

ACTION

-0.64

POSITIVE LOGITS

 Palace

0.95

 palace

0.89

 intrigue

0.88

 intrig

0.88

 Hotel

0.86

 courtyard

0.84

lain

0.79

 Theatre

0.79

 coup

0.76

 Gardens

0.75

Activations Density 0.025%

places or buildings with historical significance

No Comments

No Known Activations