INDEX

Explanations

percentages or numerical values related to increases or decreases over time

oai_token-act-pair · gpt-3.5-turbo

numerical data related to statistics and trends

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

radio

-0.70

atom

-0.65

 voic

-0.64

 grav

-0.62

aqu

-0.62

 microphones

-0.62

ãĤĮ

-0.62

 helicop

-0.62

 medi

-0.61

Ĥ¬

-0.61

POSITIVE LOGITS

 OECD

0.90

 compared

0.82

 comparison

0.80

 fiscal

0.79

0.78

0.73

adjusted

0.72

 surveyed

0.72

Activations Density 0.298%

percentages or numerical values related to increases or decreases over time

numerical data related to statistics and trends

No Comments

No Known Activations