INDEX

Explanations

information about a news organization, likely related to news content, publishers, or licensing

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

xit

-0.82

RM

-0.71

olean

-0.71

osaurus

-0.70

ruary

-0.69

ËĪ

-0.69

 Ratt

-0.68

orem

-0.68

ERSON

-0.67

posted

-0.66

POSITIVE LOGITS

bleacher

0.78

 donor

0.77

 donors

0.72

Semitism

0.71

NPR

0.70

 journalism

0.70

 affiliates

0.69

 recipients

0.66

 investigative

0.65

APD

0.64

Activations Density 0.058%

information about a news organization, likely related to news content, publishers, or licensing

No Comments

No Known Activations