INDEX

Explanations

numerical expressions, such as measurements, resolutions, and model numbers

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

TextColor

-0.74

rely

-0.68

stood

-0.66

ques

-0.66

netflix

-0.66

ournal

-0.65

inas

-0.62

loo

-0.62

unct

-0.61

raviolet

-0.60

POSITIVE LOGITS

acity

1.11

^{

0.92

acious

0.89

 Downing

0.88

aciously

0.86

Gb

0.82

0.81

uther

0.81

tons

0.78

 commandments

0.77

Activations Density 0.559%

numerical expressions, such as measurements, resolutions, and model numbers

No Comments

No Known Activations