INDEX

Explanations

mentions of the acronym "IM" followed by a number, often related to various topics and organizations

oai_token-act-pair · gpt-3.5-turbo

references to specific organizations or institutions, particularly related to "IM"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

aeus

-0.76

 elves

-0.69

 ducks

-0.68

 broccoli

-0.64

 filled

-0.63

 cleaners

-0.63

 fries

-0.61

 dyed

-0.61

 Shoals

-0.60

 stew

-0.60

POSITIVE LOGITS

IM

4.07

IME

1.96

im

1.92

ims

1.77

imo

1.54

IM

1.52

ima

1.50

AMI

1.47

AM

1.46

IF

1.44

Activations Density 0.006%

mentions of the acronym "IM" followed by a number, often related to various topics and organizations

references to specific organizations or institutions, particularly related to "IM"

No Comments

No Known Activations