INDEX

Explanations

locations specified by distance measurements

oai_token-act-pair · gpt-3.5-turbo

references to proximity or distance

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

OY

-0.91

ħĭ

-0.79

Ĥª

-0.79

uci

-0.76

may

-0.72

oy

-0.70

KC

-0.67

cour

-0.66

ardi

-0.66

LC

-0.64

POSITIVE LOGITS

 nearest

1.11

 landfall

0.81

 shore

0.79

 offending

0.75

 closest

0.75

 actual

0.74

 coastline

0.72

 peak

0.68

 baseline

0.66

 nonexistent

0.65

Activations Density 0.135%

locations specified by distance measurements

references to proximity or distance

No Comments

No Known Activations