INDEX

Explanations

the word "imp" or variations of it

oai_token-act-pair · gpt-3.5-turbo

the term "imp" as related to implications or inherently problematic situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Amend

-0.74

 Seym

-0.73

APD

-0.71

 Hath

-0.71

£ı

-0.70

 Democr

-0.69

 Munich

-0.68

 McMaster

-0.68

LIA

-0.68

Accessory

-0.67

POSITIVE LOGITS

ossible

1.25

ossibly

1.15

regn

1.08

ulse

1.08

ression

1.08

imp

1.07

romptu

1.07

icult

1.06

ressive

1.06

ocalypse

1.04

Activations Density 0.007%

the word "imp" or variations of it

the term "imp" as related to implications or inherently problematic situations

No Comments

No Known Activations