INDEX

Explanations

mentions of the word 'Tweedale'

oai_token-act-pair · gpt-3.5-turbo

references to "Tweet" or its variations, indicating a focus on social media posts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ILY

-0.81

OPA

-0.72

BILITY

-0.72

ãĥĩ

-0.72

IDS

-0.69

RAD

-0.67

 Defenders

-0.65

 Romanian

-0.65

 Valiant

-0.64

VAL

-0.64

POSITIVE LOGITS

Twe

1.29

Twe

1.02

eden

0.99

edy

0.94

ets

0.86

twe

0.86

ollen

0.85

aker

0.85

uner

0.84

enery

0.84

Activations Density 0.004%

mentions of the word 'Tweedale'

references to "Tweet" or its variations, indicating a focus on social media posts

No Comments

No Known Activations