INDEX

Explanations

This neuron is looking for words related to imperfection or flaws

oai_token-act-pair · gpt-3.5-turbo

terms related to permanence and inevitability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

anwhile

-0.85

wagen

-0.80

 guiActiveUnfocused

-0.70

CDs

-0.68

âĸ¬

-0.67

 GOODMAN

-0.66

hops

-0.65

hare

-0.65

WAYS

-0.64

ãĥ¯ãĥ³

-0.64

POSITIVE LOGITS

vious

1.13

ishable

1.13

missible

1.10

manent

1.09

iled

0.89

mented

0.89

bably

0.89

pex

0.87

redict

0.86

igen

0.85

Activations Density 0.012%

This neuron is looking for words related to imperfection or flaws

terms related to permanence and inevitability

No Comments

No Known Activations