INDEX

Explanations

phrases related to titles or headings

oai_token-act-pair · gpt-3.5-turbo

statements indicating existence or descriptions of various subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

isms

-0.75

arming

-0.73

continue

-0.72

Shape

-0.69

conn

-0.68

grow

-0.67

urden

-0.66

 resorts

-0.65

fe

-0.65

pite

-0.65

POSITIVE LOGITS

 located

0.90

 basically

0.90

 preceded

0.84

 supposed

0.84

 called

0.82

 essentially

0.82

 comprised

0.81

 named

0.80

 accompanied

0.78

 billed

0.77

Activations Density 0.440%

phrases related to titles or headings

statements indicating existence or descriptions of various subjects

No Comments

No Known Activations