INDEX

Explanations

phrases related to illegal or problematic actions and consequences

oai_token-act-pair · gpt-3.5-turbo

statements or claims about various scenarios, often using the verb "is" to assert conditions or attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Awakens

-0.66

onday

-0.62

 Telesc

-0.62

 Came

-0.61

ãĤ©

-0.59

 Scroll

-0.58

sburg

-0.57

kson

-0.57

peg

-0.57

 Brow

-0.56

POSITIVE LOGITS

 prohibited

1.15

 commonplace

1.12

 often

1.11

 usually

1.11

rael

1.11

 commonly

1.07

 rarely

1.04

 outlawed

1.03

 frowned

1.02

 generally

1.02

Activations Density 0.292%

phrases related to illegal or problematic actions and consequences

statements or claims about various scenarios, often using the verb "is" to assert conditions or attributes

No Comments

No Known Activations