INDEX

Explanations

words related to the external or physical aspects of objects or situations

oai_token-act-pair · gpt-3.5-turbo

references to “outside” and “inside” locations or conditions

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

iji

-0.92

 Reincarn

-0.76

 Rampage

-0.76

cycles

-0.75

ammy

-0.75

zl

-0.74

ruary

-0.72

enegger

-0.72

lished

-0.72

lication

-0.69

POSITIVE LOGITS

tip

0.74

 underside

0.74

 ridge

0.74

 observer

0.72

most

0.71

 workings

0.70

 wing

0.68

 panel

0.67

 flap

0.66

 attractiveness

0.64

Activations Density 0.038%

words related to the external or physical aspects of objects or situations

references to “outside” and “inside” locations or conditions

No Comments

No Known Activations