INDEX

Explanations

reasons or explanations for specific situations

oai_token-act-pair · gpt-3.5-turbo

mentions of "reasons" 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

yss

-0.70

oba

-0.67

enged

-0.67

ibaba

-0.66

ILA

-0.64

boro

-0.62

 tatt

-0.60

 likeness

-0.59

aila

-0.59

ascus

-0.59

POSITIVE LOGITS

 reasons

1.06

abl

0.94

 unrelated

0.90

why

0.87

 Reasons

0.82

Â¶

0.81

asons

0.76

 alone

0.75

why

0.74

Reason

0.74

Activations Density 0.028%

reasons or explanations for specific situations

mentions of "reasons" in various contexts

No Comments

No Known Activations