INDEX

Explanations

expert opinions or analysis in a text

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to experts and analyses in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

ò

-0.82

EStream

-0.74

ibrary

-0.74

ĪĴ

-0.73

ãĤ¦ãĤ¹

-0.73

ñ

-0.73

vertisement

-0.70

Interstitial

-0.70

History

-0.70

Reloaded

-0.69

POSITIVE LOGITS

 predict

1.24

 warn

1.17

 agree

1.17

 estimate

1.13

 speculate

1.11

 reckon

1.09

say

1.04

 caution

1.02

 disagree

1.01

 cautioned

1.00

Activations Density 0.110%

expert opinions or analysis in a text

references to experts and analyses in various contexts

No Comments

No Known Activations