INDEX

Explanations

words related to promoting or promoting actions

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

*/(

-0.82

psons

-0.80

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.76

ANG

-0.75

assian

-0.75

displayText

-0.74

 partName

-0.71

 Detected

-0.71

fuck

-0.70

ãĤ«

-0.68

POSITIVE LOGITS

 abstinence

0.96

 atheism

0.91

 separat

0.90

 boycot

0.83

 virtues

0.81

 tolerance

0.80

 wellness

0.80

 unity

0.80

 patriotism

0.80

 authenticity

0.79

Activations Density 0.082%

words related to promoting or promoting actions

No Comments

No Known Activations