INDEX

Explanations

technical computer-related terms

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 decomp

-0.78

 JPEG

-0.76

 fortun

-0.71

 photoc

-0.70

 scrut

-0.68

 pyramid

-0.68

 silhou

-0.67

 scrim

-0.67

 hurd

-0.67

jog

-0.66

POSITIVE LOGITS

ï¸ı

1.26

1.19

ski

1.17

ship

1.16

tre

1.05

iversary

1.04

tal

1.02

sure

0.95

mental

0.94

tsy

0.90

Activations Density 0.987%

technical computer-related terms

No Comments

No Known Activations