INDEX

Explanations

references to terms related to a specific acronym or abbreviation (e.g., "PD")

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "PD" followed by numerical values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

bourg

-0.88

ansas

-0.77

itud

-0.71

ppo

-0.67

hold

-0.66

natureconservancy

-0.66

ught

-0.65

votes

-0.65

fee

-0.64

tty

-0.64

POSITIVE LOGITS

ouble

0.86

DR

0.79

GF

0.76

PD

0.75

 Constable

0.74

ective

0.73

iamond

0.72

oodle

0.72

eeper

0.71

Activations Density 0.041%

references to terms related to a specific acronym or abbreviation (e.g., "PD")

references to the abbreviation "PD" followed by numerical values

No Comments

No Known Activations