INDEX

Explanations

words related to investigation or scrutiny

oai_token-act-pair · gpt-3.5-turbo

words related to actions of investigation, alteration, or restriction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Sierra

-0.76

 VIDEOS

-0.72

 ability

-0.63

 formation

-0.63

 formations

-0.61

dos

-0.61

 snippet

-0.60

ggles

-0.59

 spont

-0.59

 constitutes

-0.58

POSITIVE LOGITS

rolled

1.03

idered

0.98

orthy

0.97

icent

0.96

ivated

0.92

omorphic

0.92

itored

0.92

ivable

0.91

igated

0.91

rozen

0.91

Activations Density 0.329%

words related to investigation or scrutiny

words related to actions of investigation, alteration, or restriction

No Comments

No Known Activations