INDEX

Explanations

words related to instructions or commands

oai_token-act-pair · gpt-3.5-turbo

phrases related to ratings and assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

arching

-0.66

theless

-0.64

 brisk

-0.58

stretched

-0.58

Ã¡s

-0.57

Ô

-0.57

ãĥ£

-0.56

ulner

-0.56

 Sask

-0.55

anchester

-0.55

POSITIVE LOGITS

pedia

0.81

 Anniversary

0.61

 Worlds

0.58

monkey

0.57

 Despair

0.56

hood

0.55

 trace

0.55

 lockout

0.54

 Annotations

0.53

 Oaks

0.51

Activations Density 0.759%

words related to instructions or commands

phrases related to ratings and assessments

No Comments

No Known Activations