INDEX

Explanations

references to history and following a predecessor's path

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ouf

-0.61

 typo

-0.61

 mismatch

-0.57

atron

-0.57

iza

-0.55

idays

-0.55

 Limit

-0.54

 matchup

-0.53

 ingested

-0.53

 desktop

-0.52

POSITIVE LOGITS

ordinate

0.82

Emer

0.75

animate

0.74

same

0.72

DEP

0.71

ä½ľ

0.71

 footsteps

0.71

à¦

0.71

HER

0.70

Activations Density 0.066%

references to history and following a predecessor's path

No Comments

No Known Activations