INDEX

Explanations

names of places or people related to a particular event

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

istics

-0.76

ocl

-0.74

arily

-0.71

 GOODMAN

-0.71

 Mehran

-0.70

istical

-0.69

ifact

-0.69

uania

-0.68

igating

-0.68

istically

-0.68

POSITIVE LOGITS

TON

0.90

ff

0.89

tan

0.88

xon

0.86

ppo

0.84

lli

0.84

ption

0.83

ira

0.81

trop

0.80

crow

0.75

Activations Density 0.054%

names of places or people related to a particular event

No Comments

No Known Activations