INDEX

Explanations

words related to making choices or decisions

oai_token-act-pair · gpt-3.5-turbo

instances of selectivity 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ASC

-0.81

ijk

-0.78

FORE

-0.77

zl

-0.77

ovember

-0.76

ruck

-0.75

Downloadha

-0.72

Jenn

-0.71

Â¯Â¯

-0.71

CTV

-0.71

POSITIVE LOGITS

 selective

0.94

 inhibitors

0.89

 impulse

0.89

 pressures

0.87

 choice

0.87

 breeding

0.85

allery

0.84

 sensitivity

0.80

 serotonin

0.79

 exclus

0.77

Activations Density 0.013%

words related to making choices or decisions

instances of selectivity in various contexts

No Comments

No Known Activations