INDEX

Explanations

phrases related to fears and warnings

oai_token-act-pair · gpt-3.5-turbo

phrases that express existential or philosophical concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

TPS

-0.79

NM

-0.79

CF

-0.78

 Cosponsors

-0.74

 Proced

-0.71

CB

-0.70

related

-0.69

reporting

-0.68

seys

-0.67

LOG

-0.67

POSITIVE LOGITS

thy

1.11

 ..."

1.09

 âĢ¦"

1.09

 fools

1.07

 thou

1.07

nig

1.05

 evil

1.02

 shalt

0.98

 mankind

0.98

 tyranny

0.96

Activations Density 0.462%

phrases related to fears and warnings

phrases that express existential or philosophical concepts

No Comments

No Known Activations