INDEX

Explanations

measurements, specifically related to depth and monetary values

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

ctor

-0.74

Synopsis

-0.71

scripts

-0.67

Commerce

-0.67

values

-0.67

inf

-0.63

Keys

-0.61

hran

-0.58

formation

-0.58

gal

-0.57

POSITIVE LOGITS

 increments

1.20

 apiece

1.18

+.

0.90

 altogether

0.81

 thereafter

0.79

 total

0.79

+,

0.77

/$

0.75

 deducted

0.74

 rounding

0.73

Activations Density 0.403%

measurements, specifically related to depth and monetary values

No Comments

No Known Activations