INDEX

Explanations

technical information and terms related to technology

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

isy

-0.74

jad

-0.69

othe

-0.61

lie

-0.59

odes

-0.59

ire

-0.59

 imaginable

-0.57

heartedly

-0.57

nat

-0.56

 opin

-0.55

POSITIVE LOGITS

 which

1.27

which

1.12

aka

1.11

 whose

1.08

 wherein

1.05

 whereby

1.05

 where

0.98

 whence

0.98

whose

0.94

Inc

0.89

Activations Density 2.354%

technical information and terms related to technology

No Comments

No Known Activations