INDEX

Explanations

descriptions of being completely immersed or surrounded by a particular experience or environment

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

alias

-0.74

rer

-0.72

²¾

-0.71

boxing

-0.71

NPR

-0.70

align

-0.68

arily

-0.67

Ĥİ

-0.67

taker

-0.66

aligned

-0.65

POSITIVE LOGITS

 adrenaline

1.04

 fumes

0.97

 molten

0.92

 rubble

0.91

 dirt

0.86

 fluids

0.86

 fresh

0.86

 excitement

0.85

mud

0.85

 flames

0.84

Activations Density 1.490%

descriptions of being completely immersed or surrounded by a particular experience or environment

No Comments

No Known Activations