INDEX

Explanations

indications of a call to action

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate actions or recommendations to be taken

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

person

-0.77

Sold

-0.76

hent

-0.70

raped

-0.68

listed

-0.67

rn

-0.64

DERR

-0.59

 Beats

-0.59

backed

-0.58

volent

-0.58

POSITIVE LOGITS

 celebrate

1.16

 revisit

1.15

 settle

1.06

 conserve

1.02

 refresh

0.99

 revise

0.98

 retire

0.96

 resolve

0.92

 consolidate

0.92

 introduce

0.92

Activations Density 0.071%

indications of a call to action

phrases that indicate actions or recommendations to be taken

No Comments

No Known Activations