INDEX

Explanations

words related to explanations or justifications

oai_token-act-pair · gpt-3.5-turbo

phrases indicating justification or explanation

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

 Riders

-0.73

 helicop

-0.69

pocket

-0.68

 pocket

-0.66

 needle

-0.64

KY

-0.64

agra

-0.62

 stocking

-0.61

 Rider

-0.60

 Saskatchewan

-0.60

POSITIVE LOGITS

abl

1.02

 reasons

0.92

 sake

0.85

asons

0.79

reason

0.77

mpeg

0.72

 unrelated

0.71

 Reasons

0.70

 inexplicable

0.69

neum

0.69

Activations Density 0.016%

words related to explanations or justifications

phrases indicating justification or explanation

No Comments

No Known Activations