INDEX

Explanations

words related to disparities and discrepancies

oai_token-act-pair · gpt-3.5-turbo

terms related to inequality and differences between groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 sworn

-0.69

cend

-0.68

 guided

-0.66

safe

-0.63

 bark

-0.63

 Kiss

-0.63

 Commands

-0.63

assad

-0.63

help

-0.62

Tag

-0.62

POSITIVE LOGITS

 disparity

3.01

 discrepancy

2.88

 discrepancies

2.68

 disparities

2.54

 imbalance

2.39

 mismatch

2.09

 inconsistencies

2.07

 inconsistency

2.00

 disproportion

1.96

 discrep

1.92

Activations Density 0.035%

words related to disparities and discrepancies

terms related to inequality and differences between groups

No Comments

No Known Activations