INDEX

Explanations

terms related to reactions or replies

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "responses."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enture

-0.76

andal

-0.70

ijah

-0.66

 Skydragon

-0.66

 illegally

-0.65

board

-0.63

uke

-0.63

akin

-0.61

plane

-0.60

 dynasty

-0.60

POSITIVE LOGITS

 responses

3.59

 replies

2.37

 reactions

2.30

 answers

2.11

 response

1.77

response

1.67

 submissions

1.60

 Respons

1.59

 responds

1.56

 Answers

1.49

Activations Density 0.012%

terms related to reactions or replies

occurrences of the word "responses."

No Comments

No Known Activations