INDEX

Explanations

words related to contrast or contradiction, where the reality is different from the initial impression or belief

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.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

pol

-0.69

Pont

-0.60

 itch

-0.59

 finalists

-0.58

itialized

-0.58

chairs

-0.58

 Lethal

-0.56

Americ

-0.56

 Instr

-0.55

usable

-0.55

POSITIVE LOGITS

ãĥł

0.69

{:

0.68

 neither

0.65

estro

0.64

 Fraud

0.63

rimination

0.63

é¾į

0.62

aris

0.61

wagen

0.60

å·

0.59

Activations Density 0.084%

words related to contrast or contradiction, where the reality is different from the initial impression or belief

No Comments

No Known Activations