INDEX

Explanations

references to legal sentences and their associated terms

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

BLIC

-0.84

sie

-0.78

wagen

-0.75

ergy

-0.72

sonian

-0.71

chio

-0.70

endor

-0.69

drive

-0.67

CRE

-0.67

ãĤ¸

-0.65

POSITIVE LOGITS

 sentences

1.20

 sentence

1.05

 sentencing

1.04

 probation

0.91

 Guantanamo

0.88

 convict

0.87

 imposed

0.87

 prison

0.85

 incarcer

0.85

 imprisonment

0.84

Activations Density 0.037%

references to legal sentences and their associated terms

No Comments

No Known Activations