INDEX

Explanations

mentions of natural disasters, particularly earthquakes

oai_token-act-pair · gpt-3.5-turbo

references to earthquakes and their magnitudes

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

skirts

-0.96

atives

-0.79

 Purs

-0.73

secut

-0.73

 Parenthood

-0.72

orney

-0.69

chwitz

-0.68

zsche

-0.68

racuse

-0.66

leeve

-0.65

POSITIVE LOGITS

 earthquake

0.98

 rupture

0.98

 magnitude

0.92

hower

0.90

 quake

0.87

 shook

0.80

 Earthquake

0.80

 struck

0.79

 swarm

0.79

 hazard

0.79

Activations Density 0.046%

mentions of natural disasters, particularly earthquakes

references to earthquakes and their magnitudes

No Comments

No Known Activations