INDEX

Explanations

words related to identifying or locating specific details or points

oai_token-act-pair · gpt-3.5-turbo

phrases related to identifying or determining specific information or locations

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

perty

-0.82

RAW

-0.72

wife

-0.71

rentice

-0.70

annis

-0.70

egal

-0.69

wives

-0.68

boys

-0.67

tha

-0.67

ammy

-0.67

POSITIVE LOGITS

 pinpoint

1.25

ibly

0.88

pell

0.79

ãĤ©

0.75

pelled

0.75

ilated

0.71

PsyNetMessage

0.71

 accuracy

0.70

pelling

0.70

 pigeon

0.70

Activations Density 0.016%

words related to identifying or locating specific details or points

phrases related to identifying or determining specific information or locations

No Comments

No Known Activations