INDEX

Explanations

words related to physical medical conditions starting with 'ar', 'ry', 'ol', 'ed', 'i', and 'ell'

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

yip

-0.90

uala

-0.74

 Mania

-0.61

ibrary

-0.61

©¶æ

-0.61

Obi

-0.60

 Galile

-0.60

 constitu

-0.60

 Celebrity

-0.59

 gears

-0.58

POSITIVE LOGITS

enment

1.10

vable

0.83

ement

0.82

ened

0.74

aration

0.74

rill

0.73

ment

0.73

bsite

0.72

eling

0.72

achment

0.72

Activations Density 0.067%

words related to physical medical conditions starting with 'ar', 'ry', 'ol', 'ed', 'i', and 'ell'

No Comments

No Known Activations