INDEX

Explanations

phrases related to personal beliefs and actions

oai_token-act-pair · gpt-3.5-turbo

references to personal beliefs and skepticism towards authority

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

 elusive

-0.62

 uncanny

-0.61

 unofficial

-0.61

 unforeseen

-0.60

 imaginable

-0.60

 Prec

-0.60

Interested

-0.59

Hoy

-0.59

 Stead

-0.58

 overriding

-0.58

POSITIVE LOGITS

 invested

0.76

elf

0.75

selves

0.74

self

0.70

own

0.67

enery

0.67

pees

0.66

 touched

0.66

tein

0.65

 preached

0.65

Activations Density 0.371%

phrases related to personal beliefs and actions

references to personal beliefs and skepticism towards authority

No Comments

No Known Activations