INDEX

Explanations

arguments or discussions related to a specific topic, possibly involving pros and cons or multiple viewpoints

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

 nearby

-0.61

orer

-0.60

ades

-0.60

velt

-0.60

sold

-0.59

enium

-0.59

 horrified

-0.59

 proud

-0.58

nor

-0.58

wick

-0.55

POSITIVE LOGITS

 Firstly

1.41

 namely

1.25

Firstly

1.11

âĹı

1.04

âĢ¢

0.91

%:

0.90

viz

0.86

Â·

0.84

 includ

0.83

:-

0.82

Activations Density 2.701%

arguments or discussions related to a specific topic, possibly involving pros and cons or multiple viewpoints

No Comments

No Known Activations