INDEX

Explanations

words related to high intensity or extreme characteristics, often accompanied by positive connotations

oai_token-act-pair · gpt-3.5-turbo

the word "quite" 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ULTS

-0.84

 DRAG

-0.77

owe

-0.75

":"/

-0.74

ppo

-0.74

ults

-0.73

wa

-0.71

orer

-0.68

Reviewer

-0.67

=/

-0.67

POSITIVE LOGITS

 frankly

1.09

 literally

1.01

 possibly

0.95

 alright

0.80

 tame

0.80

 literal

0.77

 amusing

0.77

 different

0.74

 honestly

0.72

 capable

0.71

Activations Density 0.016%

words related to high intensity or extreme characteristics, often accompanied by positive connotations

the word "quite" in various contexts

No Comments

No Known Activations