INDEX

Explanations

time-related information such as dates, times, locations, and specific details in posts or profiles

oai_token-act-pair · gpt-3.5-turbo

timestamps and editing details in posts

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

 pestic

-0.77

channelAvailability

-0.72

 nuns

-0.72

 missionaries

-0.71

fell

-0.71

 Palestin

-0.71

senal

-0.67

gettable

-0.67

 Episcopal

-0.67

 Boko

-0.66

POSITIVE LOGITS

 Posts

0.79

ertodd

0.78

 Edited

0.71

gres

0.71

azo

0.70

LW

0.70

Thread

0.69

 gmaxwell

0.68

estamp

0.68

UTC

0.67

Activations Density 0.036%

time-related information such as dates, times, locations, and specific details in posts or profiles

timestamps and editing details in posts

No Comments

No Known Activations