INDEX

Explanations

mentions of the word "Dakota."

oai_token-act-pair · gpt-3.5-turbo

references to the term "ota" in various contexts

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

def

-0.62

hip

-0.61

 Dynam

-0.60

 divided

-0.59

 defamation

-0.59

 jihadist

-0.59

 Warn

-0.59

 satire

-0.59

 warn

-0.58

 Comedy

-0.57

POSITIVE LOGITS

ota

4.87

OTA

2.69

otes

1.47

ote

1.37

ot

1.35

otta

1.29

otal

1.22

ots

1.20

oda

1.13

orum

1.12

Activations Density 0.012%

mentions of the word "Dakota."

references to the term "ota" in various contexts

No Comments

No Known Activations