INDEX

Explanations

adjectives and adverbs that express degree or certainty, as well as conjunctions

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

Negative Logits

estial

-0.70

abytes

-0.68

ypes

-0.67

 capacities

-0.64

 denominations

-0.62

idences

-0.61

amins

-0.61

endars

-0.59

 inconven

-0.59

 Creatures

-0.58

POSITIVE LOGITS

 nutshell

0.84

 moot

0.79

 arises

0.78

ACTED

0.74

 Atkinson

0.69

 boils

0.67

 motivating

0.64

escap

0.63

 ling

0.62

 Murray

0.61

Activations Density 0.241%

adjectives and adverbs that express degree or certainty, as well as conjunctions

No Known Activations