INDEX

Explanations

words related to criticism or negative evaluation

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

maid

-0.82

pring

-0.70

ulative

-0.70

 Springs

-0.69

yip

-0.69

osi

-0.67

roe

-0.67

hao

-0.67

yer

-0.65

bourg

-0.65

POSITIVE LOGITS

 disregard

0.83

 reliant

0.82

heartedly

0.80

 disreg

0.79

 unrelated

0.76

 devoid

0.72

 dependent

0.71

 antit

0.71

 unaware

0.70

 revamped

0.69

Activations Density 0.326%

words related to criticism or negative evaluation

No Comments

No Known Activations