INDEX

Explanations

time-related information such as durations, periods or points in time

oai_token-act-pair · gpt-3.5-turbo

time periods related to experiences or events

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

quote

-0.73

soDeliveryDate

-0.72

cause

-0.71

JV

-0.71

Pri

-0.68

Area

-0.65

MpServer

-0.65

ashtra

-0.64

quickShipAvailable

-0.64

 Medals

-0.62

POSITIVE LOGITS

 hiatus

0.98

cation

0.88

 exhausting

0.79

 adjourn

0.77

iatus

0.74

 stal

0.73

 setbacks

0.71

 interruption

0.70

pection

0.70

dding

0.70

Activations Density 0.170%

time-related information such as durations, periods or points in time

time periods related to experiences or events

No Comments

No Known Activations