INDEX

Explanations

phrases related to analysis or examination

oai_token-act-pair · gpt-3.5-turbo

the term "apologist" and its variations, indicating discussions about defending or justifying certain viewpoints or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Hurricanes

-0.84

enegger

-0.73

 Instruments

-0.73

 Brothers

-0.72

sword

-0.70

WHERE

-0.69

ãĤ¬

-0.68

 Sands

-0.68

 Dolphins

-0.67

 Apprentice

-0.67

POSITIVE LOGITS

ollo

1.14

otrop

1.12

osite

1.08

alach

1.08

ocry

1.04

arently

1.04

reci

1.00

orters

0.99

reciation

0.98

onent

0.98

Activations Density 0.012%

phrases related to analysis or examination

the term "apologist" and its variations, indicating discussions about defending or justifying certain viewpoints or actions

No Comments

No Known Activations