INDEX

Explanations

names of locations or organizations along with the year they were founded or started

oai_token-act-pair · gpt-3.5-turbo

information related to the founding years and historical details of entities and events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

chance

-0.64

 Journals

-0.63

aukee

-0.63

 dstg

-0.62

Especially

-0.62

yip

-0.61

NULL

-0.61

abus

-0.61

 Possible

-0.60

ubi

-0.60

POSITIVE LOGITS

 subsequently

1.34

 morphed

1.18

 became

1.17

 transitioned

1.16

 remained

1.16

 lasted

1.15

sequently

1.12

 eventually

1.11

 quickly

1.10

 grew

1.09

Activations Density 0.527%

names of locations or organizations along with the year they were founded or started

information related to the founding years and historical details of entities and events

No Comments

No Known Activations