INDEX

Explanations

words related to debate or argument

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'f' in various contexts

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

 paraly

-0.73

Plex

-0.71

 Hole

-0.70

HAEL

-0.70

 Archdemon

-0.69

 Lizard

-0.67

 Crusade

-0.66

 Madness

-0.66

DOWN

-0.65

channelAvailability

-0.65

POSITIVE LOGITS

iddling

1.35

itted

1.17

ingers

1.13

athom

1.13

udging

1.10

rozen

1.09

udge

1.08

enn

1.08

airs

1.07

iddler

1.07

Activations Density 0.029%

words related to debate or argument

occurrences of the letter 'f' in various contexts

No Comments

No Known Activations