INDEX

Explanations

words related to medical conditions or technological components

oai_token-act-pair · gpt-3.5-turbo

terms related to medical conditions and limitations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

erie

-0.79

ģ

-0.74

boss

-0.71

way

-0.68

Ð¾

-0.66

ways

-0.64

opard

-0.63

RECT

-0.63

stration

-0.63

Ģ

-0.63

POSITIVE LOGITS

kinson

0.84

acebook

0.75

itance

0.75

steen

0.70

ï¸

0.69

uitous

0.69

strap

0.69

heric

0.69

ites

0.68

ensable

0.68

Activations Density 0.055%

words related to medical conditions or technological components

terms related to medical conditions and limitations

No Comments

No Known Activations