INDEX

Explanations

phrases related to events being held or taking place

oai_token-act-pair · gpt-3.5-turbo

references to events or meetings being scheduled

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

 deceive

-0.74

include

-0.69

clerosis

-0.68

naire

-0.65

ately

-0.65

naires

-0.63

 Emails

-0.62

inx

-0.62

ettel

-0.60

 adventurer

-0.60

POSITIVE LOGITS

 accountable

0.91

 hostage

0.82

 Held

0.81

ãĥĺ

0.78

zona

0.78

itud

0.78

 sway

0.78

 captive

0.77

held

0.75

hold

0.74

Activations Density 0.031%

phrases related to events being held or taking place

references to events or meetings being scheduled

No Comments

No Known Activations