INDEX

Explanations

references to singular entities or concepts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inem

-0.68

primary

-0.64

current

-0.63

apse

-0.62

ãĤ¼ãĤ¦ãĤ¹

-0.60

arden

-0.60

mun

-0.60

Ļ½

-0.59

accompan

-0.59

access

-0.58

POSITIVE LOGITS

 mention

0.84

icable

0.83

 bothering

0.76

LESS

0.76

 ounce

0.73

nor

0.73

 kidding

0.73

 dime

0.71

ches

0.69

icably

0.68

Activations Density 0.072%

references to singular entities or concepts

No Comments

No Known Activations