INDEX

Explanations

terms related to investigations and evidence

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

odor

-0.74

ldom

-0.73

liest

-0.70

cius

-0.68

 catcher

-0.66

pei

-0.65

Ely

-0.64

ipe

-0.64

enos

-0.62

ertodd

-0.61

POSITIVE LOGITS

ance

0.99

 refinement

0.86

 refine

0.78

ado

0.78

 refining

0.77

 compl

0.76

noon

0.76

 subdiv

0.75

 largeDownload

0.73

rawdownloadcloneembedreportprint

0.73

Activations Density 0.028%

terms related to investigations and evidence

No Comments

No Known Activations