INDEX

Explanations

adjectives for different kinds of things

oai_token-act-pair · gpt-3.5-turbo

terms indicating various categories or types of items or situations

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

edia

-0.95

 presidency

-0.73

ahime

-0.72

opsis

-0.72

ulhu

-0.69

eka

-0.67

aeper

-0.67

instein

-0.67

 rapist

-0.66

gary

-0.66

POSITIVE LOGITS

 kinds

0.77

 imaginable

0.77

 goodies

0.77

hell

0.77

hots

0.71

 varied

0.71

 different

0.69

itionally

0.69

alities

0.69

 things

0.68

Activations Density 0.021%

adjectives for different kinds of things

terms indicating various categories or types of items or situations

No Comments

No Known Activations