INDEX

Explanations

names or terms related to the word "Man" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual or character named "Man."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ä½ľ

-0.88

çĦ

-0.83

èª

-0.82

PsyNetMessage

-0.76

ENC

-0.76

èĥ

-0.75

hower

-0.74

çĶŁ

-0.73

ĸļ

-0.73

ç«

-0.72

POSITIVE LOGITS

uscript

1.50

oeuv

1.20

agements

1.17

agers

1.16

fred

1.09

ufact

1.05

hunt

1.03

nered

0.98

iac

0.96

ifest

0.96

Activations Density 0.020%

names or terms related to the word "Man" followed by a number

references to a specific individual or character named "Man."

No Comments

No Known Activations